I'm crossing over into the Digital Humanities lately and finding some great resources in the blogosphere. Via The Stoa Consortium and GIS for Archaeology and CRM I've been clued in to Digging Digitally and a good collection of links concerning data, metadata, and the platonic semantic cage.
The GIS industry/community has generally been solidly structure-first, as opposed to data-first. It's necessarily different in the Humanities, History and Archaeology in particular. It may take years before found artifacts are classified in the same authoritative way that we classify streets and parcels. Even then, the classification of artifacts is always open to reinterpretation as future discoveries are made. In the meanwhile, researchers need useful, plausible labels and descriptors for artifacts. It looks to me like the search for a middle way between ontology and folksonomy is well underway in the Digital Humanities.